Oxidation States and Their Assignment

Oxidation state assignment is the indispensable first step in balancing any redox equation by the half-reaction method — you cannot identify which species is oxidized or reduced, nor how many electrons are transferred, without first tracking oxidation numbers atom by atom. It underlies the construction of Latimer Diagrams, which arrange a element's oxidation states by standard reduction potential to instantly reveal whether an intermediate species is prone to disproportionation. In coordination and organometallic chemistry, correctly assigning the metal oxidation state (via the L/X donor-pair method) is the entry point for predicting d-electron count, which in turn governs geometry preferences, magnetism, and catalytic reactivity — a Ni(0) d¹⁰ center behaves utterly differently from a Ni(IV) d⁶ center in catalysis, even though both might be drawn from the "same" ligand set under different (and sometimes ambiguous) electron-counting conventions, as the dithiolene example above illustrates. This matters directly in homogeneous catalysis: oxidative addition and reductive elimination steps in cross-coupling catalysis (Pd(0)/Pd(II) cycles in Suzuki and Negishi couplings) are literally named for the oxidation-state change they induce at the metal center. In bioinorganic chemistry, tracking oxidation states is central to understanding metalloenzyme mechanisms — the iron centers in cytochrome P450 cycle through Fe(III), Fe(IV) (as a ferryl-oxo, ["Compound I"]), and back, and correctly assigning these states (using exactly the peroxide-vs-superoxide logic worked through above) is essential to interpreting Mössbauer and EPR spectra used to characterize reaction intermediates. Photosystem II's oxygen-evolving complex cycles a tetramanganese cluster through a sequence of formal Mn oxidation states (the Kok cycle S-states) en route to water oxidation, and geochemists use iron and manganese oxidation-state ratios (Fe²⁺/Fe³⁺, Mn(II)/Mn(IV)) in minerals and sediments as proxies for ancient atmospheric oxygen levels, directly connecting this bookkeeping formalism to the deep history of Earth's atmosphere.

:::quiz[Using the ionic/donor-pair method, what is the oxidation state of chromium in [Cr(CO)₆]?]

• +6
• +3
• 0 ✓
• −6

Ask the tutor: How do I decide, for an unfamiliar ligand, whether it should be treated as L-type or X-type when assigning a metal's oxidation state?

Start from a Lewis structure — the same tool used for shapes and formal charges — but read it with a different rule. In formal charge, every bonding pair is split 50/50 between the two bonded atoms, regardless of which atom "deserves" it more. In oxidation-state bookkeeping the opposite convention is used: every bonding pair is assigned entirely to the more electronegative atom. Lone pairs already belong to their atom and are untouched. Once every bonding electron has been reassigned this way, the oxidation number of an atom is:

This is precisely how a monatomic ion's charge is generated by full ionic dissociation, extended to covalent and coordinate bonds by pretending they too fully ionize toward the more electronegative atom. For a homonuclear bond (identical electronegativities, as in Cl–Cl or C–C), the pair is split evenly since neither atom is "more" electronegative — this is why atoms in their elemental form, and identical atoms bonded to each other, always contribute 0 net oxidation-state shift from that bond. Consider ethanol, CH₃CH₂OH, redrawn bond-by-bond. Every C–H bond assigns both electrons to carbon (C is more electronegative than H, ΔEN ≈ 0.4), so each C–H bond contributes −1 to that carbon. The C–C bond is homonuclear and contributes 0 to both carbons. The C–O bond assigns both electrons to oxygen (O far more electronegative), contributing +1 to carbon and −1 to oxygen. The O–H bond assigns both electrons to oxygen, contributing −1 to oxygen and +1 to hydrogen. Working through both carbons, oxygen, and all six hydrogens reproduces the familiar result: the terminal CH₃ carbon is −3, the CH₂OH carbon is −1, oxygen is −2, and every hydrogen is +1 — summing to zero, as required for a neutral molecule. For coordination complexes the same idea is applied not bond-by-bond but ligand-by-bond: each ligand is stripped away from the metal in the closed-shell (even-electron, "ionic") form it would have as a free species, and whatever charge is left on the ligand is subtracted from the overall complex charge to find the metal's oxidation state. A chloride ligand leaves as Cl⁻ (it takes the bonding pair, being far more electronegative than most metals); an ammonia ligand leaves as neutral NH₃ (it donates a lone pair but is not stripped of any additional electron, since N–H and N–metal bonds are treated by the same electronegativity logic and the nitrogen already "owns" the lone pair being donated).

:::svg[Coordination complex: donor-pair method for [Co(NH₃)₅Cl]²⁺]

The master rule, stated compactly for any atom A in a Lewis structure:

where:

• $Z_v(A)$ is the number of valence electrons in the neutral free atom of element A,
• $n_{\text{lone}}(A)$ is the number of lone pairs residing on A in the Lewis structure,
• the sum runs over every bond involving A, with $n_{\text{pair}}$ the number of electron pairs in that bond (1 for single, 2 for double, etc.),
• $\delta(A \text{ more EN})$ equals 1 if A is the more electronegative partner in that bond (A keeps the pair), 0 if the partner is more electronegative (A loses the pair), and $\tfrac12$ if the two atoms are identical (homonuclear bond, split evenly).

Reading: this is nothing more than "valence electrons the free atom started with, minus valence electrons it ends up owning after every bond is forced fully ionic." A positive oxidation state means the atom ends up owning fewer electrons than it started with (it is formally "oxidized," having lost electron density to more electronegative partners); a negative one means it ends up owning more. The second governing relationship is the conservation law that makes oxidation states useful for balancing redox equations:

where:

• the sum runs over every atom in the molecule or ion,
• $q$ is the overall charge on the species (0 for a neutral molecule, the ionic charge for a polyatomic ion).

Reading: oxidation numbers are not independent of each other — they must always sum to the total charge actually present. This single constraint is what lets you back out an unknown oxidation state (e.g., the S in S₂O₃²⁻) once every other atom's value is fixed by convention.

Example 1 (foundational): the mean oxidation state of sulfur in tetrathionate, S₄O₆²⁻. Oxygen takes its standard value of −2 (each O bonds only to S, and O is far more electronegative than S, so every oxygen wins all its bonds — no O–O or O–F links present to break the pattern). There are 6 oxygens, contributing $6 \times (-2) = -12$. The charge-sum rule requires

This is the average oxidation state over all four sulfurs, and it is not an artifact — it is a genuinely non-integer bookkeeping result because tetrathionate's structure is $^{-}\text{O}_3\text{S–S–S–SO}_3^{-}$: the two central sulfurs are bonded only to other sulfur (homonuclear, contributing 0 each, so they sit at O.S. = 0), while the two terminal sulfurs, each bonded to three oxygens and one sulfur, sit at O.S. = +5 (matching sulfonate-like sulfur). The mean of $\{0, 0, +5, +5\}$ is indeed +2.5, confirming that the simple charge-sum formula, while algebraically correct, can mask real structural inequivalence between atoms of the same element. As a limiting-case check: if all four sulfurs were structurally equivalent (as in, say, a hypothetical symmetric tetra-sulfonate), the mean value +2.5 would also be each individual atom's true value — the discrepancy here specifically flags a structure with two chemically distinct sulfur environments.

Example 2 (applied): the metal oxidation state in a peroxo-bridged dicobalt complex, $[(\text{NH}_3)_5\text{Co}(\mu\text{-O}_2)\text{Co(NH}_3)_5]^{4+}$. This complex contains a bridging peroxide ligand, O₂²⁻, in which the two oxygens are joined by a single O–O bond (not a double bond as in free O₂) — the extra two electrons relative to neutral O₂ sit as a full negative charge of −2 spread over the ligand, i.e. each oxygen is at O.S. = −1, not the usual −2, because each oxygen now shares one of its bonds with an identical oxygen atom (homonuclear, $\delta = \tfrac12$, contributing 0 for that bond) rather than losing both bonds' worth of electrons to a more electronegative partner. Concretely: each oxygen in peroxide has 2 lone pairs (4 e⁻, unchanged), one O–O bond split evenly (contributing 1 e⁻ to each oxygen, since $\delta = \tfrac12$ for a homonuclear bond), and one O–Co bond in which oxygen (EN 3.44) is far more electronegative than cobalt (EN ≈ 1.88), so oxygen keeps both electrons of that bond (2 e⁻). Total owned = 4 + 1 + 2 = 7 electrons, so O.S.(O) = 6 − 7 = −1 per oxygen, and −2 for the whole $\text{O}_2^{2-}$ unit — exactly as the "peroxide" name implies. Using the donor-pair method on the whole complex: ten NH₃ ligands (five per cobalt) are L-type, contributing 0 each. The bridging peroxide is a single X₂-type ligand contributing −2 total. Charge conservation over the whole dimer (total charge +4, two cobalt centers, assumed equivalent by the molecule's symmetry):

Both cobalt centers are Co(III), and the bridging ligand is best described as peroxide (O₂²⁻) rather than superoxide (O₂⁻) — a distinction with real physical consequence, since the O–O bond length in the peroxo form (~1.47 Å, close to H₂O₂) is measurably longer than in the superoxo form (~1.3 Å, close to O₂⁻), providing an experimental check on the electron-counting assignment.

Example 3 (exam/competition level): resolving a non-innocent-ligand ambiguity in a dithiolene complex. Consider a nominally square-planar complex $[\text{Ni}(\text{S}_2\text{C}_2\text{Ph}_2)_2]^{0}$, a bis(dithiolene) nickel complex, overall neutral. The dithiolene ligand, 1,2-diphenyl-1,2-ethylenedithiolate, can exist in at least two electronically distinct closed-shell forms: the fully reduced enedithiolate (ene-1,2-dithiolate, $[\text{S}_2\text{C}_2\text{Ph}_2]^{2-}$, with a C=C double bond and two thiolate sulfurs, each S formally −1 as in a normal thiolate) or the fully oxidized dithioketone/dithiolene radical form, in which one electron has been removed to give a ligand-centered radical anion $[\text{S}_2\text{C}_2\text{Ph}_2]^{\bullet-}$, or even the neutral dithioketone $\text{S}_2\text{C}_2\text{Ph}_2^{0}$ with a C=S-like resonance and an S–S-shortened, C–C-lengthened backbone. If both ligands in $[\text{Ni}(\text{S}_2\text{C}_2\text{Ph}_2)_2]^0$ are assumed to be closed-shell dithiolate dianions ($-2$ each), charge conservation over the complex demands

giving an unusual, high-oxidation-state Ni(IV)-$d^6$ assignment. If instead both ligands are assumed fully oxidized to the neutral dithioketone form (0 each), the same conservation law gives

a Ni(0)-$d^{10}$ assignment — a difference of four full oxidation-state units between two chemically "reasonable" ways of drawing the same Lewis structure. This is the hallmark of a non-innocent ligand: because the frontier orbitals of the dithiolene sulfur/carbon π system lie extremely close in energy to the nickel 3d orbitals, the "true" electronic structure is a heavily mixed, partially covalent combination of these limiting resonance descriptions, and no single formal oxidation state is unambiguously "correct" from the Lewis structure alone. The resolution in practice, and the pedagogical lesson, is that oxidation-state assignment always presupposes a well-defined, localized Lewis structure with clear bond orders — X-ray crystallographic bond-length data (the C–C and C–S distances within the dithiolene backbone) and magnetic measurements are used experimentally to decide which resonance form dominates, and even then the assignment is often reported as an intermediate or ambiguous oxidation state (chemists commonly write "Ni(II)/Ni(III)/Ni(IV)" as a spread, or avoid formal oxidation states altogether for such complexes). This example is the sharp limiting case that shows oxidation number assignment is a model, not a direct observable, and the model can break down when donor and acceptor orbitals become nearly degenerate.

The systematic use of oxidation numbers as a redox bookkeeping tool crystallized gradually through the late nineteenth and early twentieth centuries, as chemists moved from purely empirical "equivalent weight" reasoning toward an electronic picture of valence. A pivotal figure in connecting electron-pair bonding to formal charge assignment was Gilbert N. Lewis, whose 1916 shared-electron-pair model of the covalent bond supplied the very structures — dot diagrams with explicit bonding and lone pairs — on which oxidation-state rules are read off. Before Lewis structures existed, "oxidation number" was essentially a name for the ionic charge inferred from stoichiometry and empirical valence rules (e.g., assuming a metal in an oxide simply took the charge needed to balance O²⁻), without any explicit electron-pair justification. The rigorous electronegativity-based definition — assign each shared pair to the more electronegative atom — became possible only once Linus Pauling formalized a quantitative electronegativity scale in the 1930s, giving chemists an operational way to decide, for any bond, which atom "wins" the pair rather than relying on memorized element-specific rules alone. IUPAC later codified the modern hierarchy of rules (elements = 0, fluorine always −1, standard values for O and H with explicit stated exceptions, charge-sum conservation) into the systematic recommendations still taught today, explicitly built on the Lewis-structure-plus-electronegativity foundation rather than treating oxidation number as a primitive, unexplained integer.